TCP Selective Acknowledgment Options

Abstract

TCP Selective Acknowledgment Options Status of this MemoThis document specifies an Internet standards track protocol for the Internet community, and requests discussion and suggestions for improvements. Please refer to the current edition of the "Internet Official Protocol Standards" (STD 1) for the standardization state and status of this protocol. Distribution of this memo is unlimited. Abstract TCP may experience poor performance when multiple packets are lost from one window of data.With the limited information available from cumulative acknowledgments, a TCP sender can only learn about a single lost packet per round trip time. An aggressive sender could choose to retransmit packets early, but such retransmitted segments may have already been successfully received.A Selective Acknowledgment (SACK) mechanism, combined with a selective repeat retransmission policy, can help to overcome these limitations. The receiving TCP sends back SACK packets to the sender informing the sender of data that has been received. The sender can then retransmit only the missing data segments.This memo proposes an implementation of SACK and discusses its performance and related issues.

Understanding the performance of many TCP flows

“…Earlier ones include [14,27,30] etc. Recently, there have been significant research efforts on developing analytical models for the dynamic behavior of TCP.…”

“…In Section 4 and Section 5, we study the effect of adding random overhead and using RED gateway on TCP performance, respectively. In Section 6, we examine the performance of TCP/SACK (TCP with support for selective acknowledgments) [30] under both Drop-Tail and RED gateways. We quantify the roundtrip bias using both simulations and analysis in Section 7.…”

Improving the Performance of TCP in the Presence of Interacting UDP Flows in Ad Hoc Networks

“…In such a scenario, the TCP flows that intersect 4 with the heavy UDP flow suffer delays and drops. TCP interprets this loss to be a mark of congestion and acts accordingly [17]. In this work, we first show the extent to which fair-queuing improves TCP performance in such a scenario.…”

Is VBR a solution for an ATM LAN ?

“…This is due to the fact that there PartTwo Video over ATM was no long enough idle time between the last data packet transmitted and the retransmitted packet, and thus cells from the retransmitted packet were also discarded by the UPC. The other enhancements to TCP's retransmission algorithm proposed recently like TCP Vegas [BOP94] or the new selective acknowledgements option [MMF96] would not have significantly changed the results of our measurements as all these strategies try to retransmit the lost packet as soon as possible. In a wide area network, however, these improved retransmission algorithms would probably be much more useful.…”